Summarize Group Discussion (PTE Academic): Format, Notes and a Model Answer

Is this PTE Academic or PTE Core?

PTE Academic. Summarize Group Discussion does not appear in Pearson’s PTE Core Score Guide (January 2026) and is not one of PTE Core’s 19 task types, so if you are sitting PTE Core for Canadian immigration you will never see it, and none of these timings apply. Pearson’s PTE Academic Test Taker Score Guide (November 2024) does not list it either, which makes Pearson’s task article the most detailed public description available. If you are preparing for Core, start at the PTE Core hub — a separate test family with different task types and word limits.

Format and timings

The layout is the same every time: instructions at the top, an audio box below them, and a status bar showing how much of the discussion is left.

Pearson does not publish how many SGD items appear in a test, so treat any figure for that as unverified.

How it is scored — and what Pearson does not publish

Pearson describes this as a partial credit task scored on three criteria: content, fluency and pronunciation. Pronunciation and oral fluency follow the same standards as other PTE speaking tasks. Content is where this task differs, and Pearson lists what a high content score requires:

• Summarise the whole discussion accurately
• Give detailed information about what each speaker says, thinks and feels
• Deliver an original response, without memorised templates or pre-prepared phrases
• Use a variety of appropriate expressions
• Connect ideas with linking words such as “however”, “on the other hand” and “as a result”
• Use appropriate academic language throughout

What is not published: there is no numeric content band for this task — no 0–5 rubric — in Pearson’s article or in the November 2024 Score Guide. Any site quoting a specific SGD content scale is inventing it. Pearson does state plainly that missing a speaker, or misrepresenting their viewpoint, costs content marks, and that all three points of view are needed for full marks. Perfection is not required: hesitations and the odd “um” appear in high-scoring responses.

Note-taking for audio you cannot replay

Three minutes is far too long to hold in your head, and you hear it once. Use a speaker-based layout: one block per speaker, drawn on your note board before the audio starts.

TOPIC: ________________________________
TYPE:  lecture discussion / course challenge / assignment planning

S1 (for / against):     S2:                    S3:
-                      -                      -
feel:                  feel:                  feel:

ALL AGREE:
DISAGREE ON:

• Keywords, not sentences. Full sentences cost the seconds you need for listening.
• Speakers do not take turns in 1-2-3 order. Identify the voice first, then write in that speaker’s block.
• One position and one feeling per speaker. “Against” and “worried” are enough to build your relationship sentences later.
• Track agreement in the margins. A quick “+S2” beside a point tells you who shared it later — the connection Pearson asks you to show.
• Log a repeated point once, under whoever raised it, and note who agreed.
• Write the topic the moment you understand it, plus any decision or solution the group reaches — that is your closing sentence.

A reusable spoken structure (organise with it, don’t recite it)

1. Orientation (one sentence). Name the topic, and the kind of discussion if it is obvious: “The three speakers were discussing whether…”
2. Speaker 1. Their position, one or two details, and how they seemed to feel.
3. Speaker 2. Their position and details, then where they part company with Speaker 1.
4. Speaker 3. Their position, what they took from each of the others, and anything new they added.
5. Relationships and close (one or two sentences). What all three agreed on, what they disagreed about, and any proposal or next step.

Keep attribution explicit as you go: “the first speaker”, “the second speaker agreed up to a point, but…”, “unlike the first speaker, she felt…”. Those are the connecting expressions Pearson looks for — keep them natural rather than fixed.

Worked example: a debate on a research question

The discussion below is invented for practice; copy the note layout and the model answer.

Scenario: three speakers discuss a seminar question — should the city charge drivers to enter the centre, in order to cut air pollution?

TOPIC: congestion charge to cut air pollution?
TYPE:  debate on a research question

S1 (for):                        S2 (sceptical):                 S3 (middle):
- pollution worst nr ring road    - charge hits those who can't   - agrees pollution urgent (+S1)
- child asthma rising              afford to live near work       - agrees fairness worry real (+S2)
- poorest households breathe      - night shifts, no buses        - smaller zone, ring-fence
  worst air                      - other cities: traffic fell       money for buses
- voluntary schemes failed         inside, rose at boundary      - review after 2 yrs
- feel: frustrated                - evidence "mixed", longer trial - resets question: which
                                  - feel: worried for low-income    combination works?
                                    drivers                       - feel: optimistic trial settles it
ALL AGREE:  current approach failing; evidence from other cities matters
DISAGREE ON: whether a charge alone is fair / effective

Model spoken answer (260 words — comfortably inside the two minutes):

The three speakers were discussing whether the city should charge drivers to enter the centre in order to cut air pollution, and they disagreed about whether that would be fair and effective. The first speaker was strongly in favour. She said pollution is worst around the ring road and that child asthma cases in the centre are rising. For her the unfairness runs the other way: the households living in the dirtiest air are the ones least likely to own a car. She seemed frustrated, and argued that a decade of voluntary schemes had already failed. The second speaker was not convinced. He accepted that traffic inside the zone might fall, but said other cities show pollution simply moving to the boundary, so the evidence is mixed. His main worry was fairness in the opposite direction: a charge falls hardest on people who cannot afford to live near their work, particularly those on night shifts with no bus service. The third speaker took a middle position. He agreed with the first speaker that the problem is urgent, and with the second that a charge on its own is unfair. He suggested a smaller zone with the revenue ring-fenced for buses, reviewed after two years. He felt a trial would settle the argument, and he reframed the question as not whether to charge drivers but which combination of measures works. In the end, all three agreed the present approach is failing and that evidence from other cities matters, but they disagreed about whether a charge alone would be fair or effective.

Why it works: the topic is named first, each speaker gets a position and a feeling, and both agreement and disagreement are explicit.

The three discussion types, and how your summary changes

Pearson does not publish a separate rubric per category, so the three types are a listening aid, not a scoring scale.

If you lose a speaker’s point mid-audio

Losing one thread for 20–30 seconds is normal, and recoverable if you react deliberately.

• Keep your hand moving. Note the voice you are hearing now; do not spend the lost time trying to rebuild the last one.
• Leave a gap and come back. Write “S2: ?” and leave room. Speakers usually return to their own thread, and you can fill it in later.
• Listen for quoting. “I agree with what he said, but…” or “unlike your point about cost…” hands you content and attribution at the same time.
• Protect structure over detail — each speaker needs a position and at least one point.
• Never stop speaking. If a block is thin, cover what you have, add a group line (“all three were concerned about…”) and move on. Fluency and pronunciation are still being scored, so silence costs more.

Four mistakes that cost the most

1. Summarising the topic, not the speakers. “They discussed an academic topic” is not a summary; Pearson’s content requirements are about what each speaker says, thinks and feels.
2. Reciting a memorised script. Original wording is a stated content requirement, and pre-prepared material sits at the 0 band of other content rubrics in the Academic Score Guide.
3. Dropping a speaker. Missing or misrepresenting one viewpoint costs content marks; all three are needed for full marks.
4. Reading your notes out loud. Keywords read aloud sound like a list with no linking language; turn each block into a sentence, with “however” or “as a result” doing the connecting.

Practise it

Rehearse the format with timed speaking items on the PTE Speaking practice page, then compare the other spoken tasks on the PTE templates page to see what is unique here. Once note-taking feels automatic, sit a full mock test so you meet this task inside a real sequence, and use the PTE essay checker for feedback on the written side. Still choosing a test for migration? Read PTE vs IELTS for Australia PR, and remember the PTE Core hub is a different test family entirely.

Frequently asked questions